The article introduces new methods to estimate autocorrelation in data without needing complex calculations. These methods are robust against extreme values and clusters of high volatility, making them useful for analyzing patterns in return series. The study also questions the common practice of adjusting conventional estimators for small-sample bias.